By Anna Rose Welch, Editorial & Community Director, Advancing RNA
In every industry, there are terms we use religiously in conversation. However, if we surveyed a room of people, we’d likely find that each term has a wide range of definitions and/or expectations. In the CGT CMC space, “flexibility,” and “phase-appropriate” are two such terms. But I’d argue the term “platform” is especially worthy of greater scrutiny in the CGT manufacturing and outsourcing space today.
Of course, we can approach this term from two different perspectives. For instance, how regulators approach “platforms” and how we as CMC professionals look at our development and/or manufacturing platforms may be slightly different. (FYI, the FDA and EMA/European Commission are exploring future guidance on this topic.) But despite the fact we in CMC rely heavily on the term “manufacturing platform” in all our discussions, the platforms that we and our outsourcing partners are working on today are not quite as cut and dried as they would be in a more standardized sector of pharma.
Given that many of us are regularly evaluating CDMOs and their CGT platform offerings, it’s becoming increasingly important to parse the reality of what capabilities truly exist behind marketing-heavy discussions of “turnkey” or “plug and play” development “platforms.” No one could have a more realistic perspective of what we’re working with today than two external manufacturing directors — one on the cell therapy side and one on the in vivo gene therapy side. Enter Adam Haskett, head of external manufacturing for cell therapy company CARGO Therapeutics, and Joseph Graskemper, formerly the senior director and head of external manufacturing & supply chain at gene therapy company Intergalactic Therapeutics, now the founder and principal of Pathways Pharma Consulting.
During two Cell & Gene Collaborative peer group sessions, Haskett and Graskemper first walked us through “the 3 ‘C’s’” of CGT outsourcing,” and how they see these factors evolving in the upcoming year(s).
But we can’t just talk broadly about our need for more advanced capabilities from our outsourcing partners without closely examining the current maturity of CDMOs’ AAV-based gene therapy and LVV-based cell therapy platforms. Lucky for us, Haskett and Graskemper shared their top tips and tricks for evaluating the seaworthiness of current CGT platform offerings. In addition to revealing their go-to questions for potential CDMO partners, they also touched on the future innovations they hope will limit or eliminate the challenges facing our processes and products today.
Defining A CGT Manufacturing Platform: AAV Gene Therapy
Obviously, I can’t harp on differences in definitions without clarifying specifically how we’re approaching the concept of “the platform” in this article. Graskemper took one for the team and clearly spelled out how he defines this term and what he expects a gene therapy platform to be capable of demonstrating when he’s pressure-testing potential gene therapy CDMO partners.
“I define ‘platform’ as a set of unit operations through upstream, downstream, and drug product that has been developed using a quality-by-design, risk-based approach that results in a robust process to deliver reproducible, high-quality material throughout multiple scales of manufacture (i.e., 2L to 2000L),” he explained. “This platform must also be supported by the corresponding analytics to reproducibly demonstrate the product’s quality and have a robust scale-down model.”
That we’re using the term “platform” more and more frequently today in our dialogues with gene therapy CDMOs is a fact that both excites and surprises Graskemper. We regularly say that “time flies when we’re having fun [developing CGTs].” But the concept of a “platform” for AAV gene therapies was all but non-existent as recently as 2019-2020. At that time, Graskemper went on a CDMO roadshow to assess adherent and suspension technologies at CDMOs located all over the world, and there was not a platform to be seen (or offered).
“If you mentioned “platform” to CDMOs back in 2019, no one had one,” Graskemper admitted. “Folks were trying to work towards something, but to achieve this rudimentary platform required a lot of co-development — and good luck if you could even successfully transfer your process into it.”
Now, just a quick four years later, thanks to great investment and much more trial and error, there are multiple platform processes available at CDMOs today. However, the maturity of these platforms varies widely — especially when we consider the sheer number of serotypes at play in the AAV field today. There are CDMOs boasting platforms that have achieved 1,000 L scale today; but it’s also critical to understand how amenable that platform is to the development of other serotypes — and if it has even successfully developed those other (non-AAV9) serotypes.
As the space continues to explore innovative methods to overcome targeting and packaging limitations, a CDMO’s experience and/or ability to integrate novel capsids into their platform is another critical consideration Graskemper weighs when evaluating CDMO’s capabilities. It’s still early days in terms of how complex these novel capsids might be to manufacture at scale, let alone within a platform employing raw materials that have only demonstrated efficacy in AAV8 or AAV9 downstream purification. Many of these raw materials have not been examined in the context of a novel capsid. “You could find yourself back at the screening point with a novel capsid,” Graskemper added.
Likewise, as Graskemper reminds us, the decision to hitch your wagon to a CDMO’s platform is just as much a business development decision as it is a scientific and technical one. The sheer existence of a platform on the market provides companies with two different paths to go down: investing more heavily in R&D and scaling-up in-house or finding a partner to take the plasmid and do all the process development and scale-up externally.
However, it’s important to approach the alluring promise of “plug and play” platform development with a healthy dose of realism today. Though “turnkey” development is a sexy addition to a potential partner’s marketing materials, current platforms do not actually offer the ability to plug in a rep-cap and gene of interest without some potentially hefty optimization and unexpected headaches.
Cell Therapy Outsourcing Platforms: “Double, Double, Toil & Trouble”?
I don’t think anyone has publicly pointed out that Shakespeare’s Macbeth (the source of this subhead’s quote) was a very early prediction of the cell therapy platform development realm. But we cannot approach a cell therapy “platform” without calling attention to two essential components commonly comprising a cell therapy: lentiviral vectors (LVV) and our cells of choice — both of which demand unique considerations in terms of outsourcing platform capabilities.
Just as different AAV serotypes and novel capsid development pose specific challenges for innovators in finding the appropriate platform for their gene therapies, the LVV side of the industry is likewise challenged by our industry’s ceaseless scientific ambitions in terms of constructs and expression.
“We’re seeing a greater need from biotechs for platforms capable of handling a greater level of construct complexity and size,” Haskett said. “We have bispecific and even trispecific constructs, which are much larger constructs to fit into a single vector. This cargo complexity also demands higher expression and a greater attention to COGS — both of which were and are still limitations of first-generation CAR-T therapies.”
Haskett sees outsourcing firms addressing these challenges in a few ways, including offering the current “gold standard” of 200-liter serum-free suspension LVV process. In fact, some firms are continuing to up-the-ante, striving to achieve 500 L or even 1,000 L suspension processes. However, as Haskett cautions, we need to be careful not to conflate higher volume production with “innovation.” There is still a great need for firms to improve quality and scalability at smaller scales. In turn, it behooves biotechs to look closely at a potential CDMO’s scale-up track record. Not only is it important to consider the number of batches a partner has produced, but sponsors should also consider how seamlessly their potential partner can scale-up from a shake flask to four or five liters, to 50 liters, to 200 L, and how well they understand their various scales and interdependencies. Like Graskemper, Haskett affirmed the importance of assessing a partner’s scale-down model as, the tighter the scale-down model, the stronger the CDMO’s understanding of the at-scale platform will be.
Such attention to scale emphasizes just how innovative successful scale-up remains in the industry. While many CDMOs can claim the same process abilities, it is often difficult to compare “apples to apples” (e.g., differences in integration titer methods) at face value. In turn, firms need to dig into the finer details of a potential partner’s scale-up datasets and demonstrated technical development skillsets. Such details can separate out the leaders from what Haskett refers to as “me too” platforms.
“There are a lot of ‘me too’ platforms,” he explained. “I think the firms that are innovating are going to be the ones that really set themselves apart and continue to be the leaders in the space.” For example, he’s keenly watching the companies that are integrating perfusion technology into their suspension processes. Others are making significant investments in expression-enhancing technologies or ramping up their molecular biology expertise and infrastructure, enabling them to tailor constructs with different regulatory elements or truncate/optimize elements while maintaining functionality.
“But of all these examples, there are only a few firms who are really innovating in these areas,” Haskett admitted.
In the cell processing realms, it goes without saying that the increasing process complexity posed by different cell types and allogeneic development is making it much more difficult to arrive at a standardized one-size-fits-all platform offering. We are seeing some early standardization happening within specific CAR-T subsets; for example, there are platform instruments available that standardize T cell selection and enrichment. However, we’re not yet at a point where we’ve achieved the same level of process standardization within the “newer” cell types (i.e., TILs, NKs). Likewise, there currently is no uniform/singular solution to accommodate the different processing needs of all the different cell therapy modalities.
We also can’t overlook the reality of the industry’s development practices. As Haskett added, “This spring, I wrapped up this CDMO roadshow, and a number of the CDMOs I talked with are building out these great, large Grade B suites because they’re still getting manual, largely open processes from their clients.”
For years, one of the biggest talking points in the cell processing space has been automation and digitalization. Though there are a few examples of “workhorse” technologies out there today, these typically only accomplish one part of the process — not to mention, these can get increasingly expensive when scaling out beyond the lab. Currently, establishing a functionally closed, automated, and complete process typically requires linking together multiple separate automation technologies, all of which have their own benefits and limitations. This is not to say that some of the leaders in this space haven’t been working on add-on technologies (e.g., formulation/fill units) for their main platform instruments. But progress remains slow — especially given the increasing modality diversity.
“That Thing You Do:” What Counts as Truly “Innovative” Today
Ultimately, for Haskett, innovation in the cell therapy space today comprises three “basic” capabilities. On the lentiviral vector side, improving titer and expression while maintaining high quality will be essential for improving overall COGS. On the processing side, closed processing and automation remain high on his list of goals for the industry to realize.
But he also sees some great unmet need — and innovation potential — in analytical and digitization efforts. “Over the last few years, I think we’ve realized that analytical capabilities are where we’re really going to move the needle,” he explained. “But I don’t see a large number of outsourcing firms innovating here.”
Not only can automating and consolidating assays lower COGS, but it can also play a significant role in enabling increased product attribute understanding and reducing cell therapy turn-around times, both of which remain a significant pain point for cell therapies today. As we covered in part 1 of this series, the industry’s increasing progress toward later-stage development and the accompanying comparability needs also necessitate top-notch analytical strategies. In-line monitoring systems that provide real-time data will go a long way for building up “waterproof” data sets for demonstrating comparability. However, as Graskemper added, the integration of PAT technologies, particularly into viral vector development, will be a long-term innovation — think a five- to 10-year journey.
Overall, as was made apparent through our conversation, we shouldn’t sell short a vendor’s ability to successfully achieve the “basics,” for example reproducible scale-up and quality. In fact, as Graskemper joked, “From an AAV platform perspective, just demonstrating that you have a reproducible, robust process is what I find to be the most innovative offering right now.” Like Haskett, he also agreed that investing long-term in automation and digitization — particularly electronic batch records and materials management — can go a long way to improving overall reproducibility.
But it also comes down to biotechs more carefully approaching each provider’s platform with the end in mind. This means not only asking if a platform will match your asset’s scalability needs, but looking more closely at whether a CDMO’s platform will provide the most feasible COGS for your product.
“Some platforms aren’t going to give you the productivity that you need, particularly at a larger scale” Graskemper added. “First, it’s important to consider how robust a CDMO’s scale-down model is, because the last thing you want is to see your productivity drop 50 percent by the time you scale-up to 200 L. Secondly, I recommend that each company build out what their COGS would look like based on a CDMO platform’s promised productivity and think near and long-term about what that will ultimately cost per-patient.”
Such calculations also open the door for biotechs to establish more specific and potentially innovative contracts with CDMOs. Thanks to more broadly available capacity and post-COVID economic changes, both Haskett and Graskemper believe contracting is a critical, often overlooked arena in which firms can better navigate the challenges of CGT platform development and COGS management today.
“We no longer have to beg CDMOs to take us on as clients anymore,” Graskemper concluded. “Developers now have greater negotiation power to come up with pricing schemes in which, for example, a CDMO is paid not just for making material but for meeting specific program milestones. Overall, I think there’s a lot more room for companies to explore novel risk-sharing arrangements in their CDMO contracts.”
Continue on to Part 3 in which Haskett, Graskemper, and I continue delving into the shape risk-sharing is taking today in CDMO-CGT biotech relationships and how this can better be reflected in contracting practices.
Missed part 1? Check it out here — From Capacity to Comparability: The Shifting Onus In Cell & Gene Therapy Outsourcing
Like what you’re reading? Sign up to continue following ARW’s writings on the CGT + RNA manufacturing paradigm.